Script to reproduce years based on a model trained with random points¶

Importing¶

In [ ]:
import xarray as xr
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

from sklearn.model_selection import train_test_split
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.compose import TransformedTargetRegressor

from sklearn.ensemble import ExtraTreesRegressor
from sklearn.ensemble import BaggingRegressor

from sklearn.metrics import root_mean_squared_error as rmse

from tqdm import tqdm

import dill
import random

import salishsea_tools.viz_tools as sa_vi

Datasets Preparation¶

In [ ]:
def datasets_preparation(dataset, dataset2):
    
    drivers = np.stack([np.ravel(dataset['Temperature_(0m-15m)']),
        np.ravel(dataset['Temperature_(15m-100m)']), 
        np.ravel(dataset['Salinity_(0m-15m)']),
        np.ravel(dataset['Salinity_(15m-100m)']),
        np.ravel(dataset2['Summation_of_solar_radiation']),
        np.ravel(dataset2['Mean_wind_speed'])
        ])
    indx = np.where(~np.isnan(drivers).any(axis=0))
    drivers = drivers[:,indx[0]]

    diat = np.ravel(dataset['Diatom'])
    diat = diat[indx[0]]

    return(drivers, diat, indx)

Regressor¶

In [ ]:
def regressor (inputs, targets):
    
    inputs = inputs.transpose()
    
    # Regressor
    X_train, _, y_train, _ = train_test_split(inputs, targets, train_size=0.35)

    model = ExtraTreesRegressor()
    model = make_pipeline(StandardScaler(), model)
    regr = BaggingRegressor(model, n_estimators=12, n_jobs=4).fit(X_train, y_train)

    return (regr)

Regressor 2¶

In [ ]:
def regressor2 (inputs, targets, variable_name):

    inputs2 = inputs.transpose()

    outputs_test = regr.predict(inputs2)
   
    m = scatter_plot(targets, outputs_test, variable_name) 
    r = np.round(np.corrcoef(targets, outputs_test)[0][1],3)
    rms = rmse(targets, outputs_test)

    return (r, rms, m)

Regressor 3¶

In [ ]:
def regressor3 (inputs, targets):
  
    inputs2 = inputs.transpose()

    outputs_test = regr.predict(inputs2)
   
    # compute slope m and intercept b
    m, b = np.polyfit(targets, outputs_test, deg=1)
    
    r = np.round(np.corrcoef(targets, outputs_test)[0][1],3)
    rms = rmse(targets, outputs_test)

    return (r, rms, m)

Regressor 4¶

In [ ]:
def regressor4 (inputs, targets, variable_name):

    inputs2 = inputs.transpose()
    
    outputs = regr.predict(inputs2)

    # Post processing
    indx2 = np.full((len(diat_i.y)*len(diat_i.x)),np.nan)
    indx2[indx[0]] = outputs
    model = np.reshape(indx2,(len(diat_i.y),len(diat_i.x)))

    m = scatter_plot(targets, outputs, variable_name + str(dates[i].date())) 

    # Preparation of the dataarray 
    model = xr.DataArray(model,
        coords = {'y': diat_i.y, 'x': diat_i.x},
        dims = ['y','x'],
        attrs=dict( long_name = variable_name + "Concentration",
        units="mmol m-2"),)
                        
    plotting3(targets, model, diat_i, variable_name)

Printing¶

In [ ]:
def printing (targets, outputs, m):

    print ('The amount of data points is', outputs.size)
    print ('The slope of the best fitting line is ', np.round(m,3))
    print ('The correlation coefficient is:', np.round(np.corrcoef(targets, outputs)[0][1],3))
    print (' The mean square error is:', rmse(targets,outputs))

Scatter Plot¶

In [ ]:
def scatter_plot(targets, outputs, variable_name):

    # compute slope m and intercept b
    m, b = np.polyfit(targets, outputs, deg=1)

    printing(targets, outputs, m)

    fig, ax = plt.subplots(2, figsize=(5,10), layout='constrained')

    ax[0].scatter(targets,outputs, alpha = 0.2, s = 10)

    lims = [np.min([ax[0].get_xlim(), ax[0].get_ylim()]),
        np.max([ax[0].get_xlim(), ax[0].get_ylim()])]

    # plot fitted y = m*x + b
    ax[0].axline(xy1=(0, b), slope=m, color='r')

    ax[0].set_xlabel('targets')
    ax[0].set_ylabel('outputs')
    ax[0].set_xlim(lims)
    ax[0].set_ylim(lims)
    ax[0].set_aspect('equal')

    ax[0].plot(lims, lims,linestyle = '--',color = 'k')

    h = ax[1].hist2d(targets,outputs, bins=100, cmap='jet', 
        range=[lims,lims], cmin=0.1, norm='log')
    
    ax[1].plot(lims, lims,linestyle = '--',color = 'k')

    # plot fitted y = m*x + b
    ax[1].axline(xy1=(0, b), slope=m, color='r')

    ax[1].set_xlabel('targets')
    ax[1].set_ylabel('outputs')
    ax[1].set_aspect('equal')

    fig.colorbar(h[3],ax=ax[1], location='bottom')

    fig.suptitle(variable_name)

    plt.show()

    return (m)

Plotting¶

In [ ]:
def plotting(variable, name):

    plt.plot(years,variable, marker = '.', linestyle = '')
    plt.xlabel('Years')
    plt.ylabel(name)
    plt.show()

Plotting 2¶

In [ ]:
def plotting2(variable,title):
    
    fig, ax = plt.subplots()

    scatter= ax.scatter(dates,variable, marker='.', c=pd.DatetimeIndex(dates).month)

    ax.legend(handles=scatter.legend_elements()[0], labels=['February','March','April'])
    fig.suptitle('Daily ' + title + ' (15 Feb - 30 Apr)')
    
    fig.show()

Plotting 3¶

In [ ]:
def plotting3(targets, model, variable, variable_name):

    fig, ax = plt.subplots(2,2, figsize = (10,15))

    cmap = plt.get_cmap('cubehelix')
    cmap.set_bad('gray')

    variable.plot(ax=ax[0,0], cmap=cmap, vmin = targets.min(), vmax =targets.max(), cbar_kwargs={'label': variable_name + ' Concentration  [mmol m-2]'})
    model.plot(ax=ax[0,1], cmap=cmap, vmin = targets.min(), vmax = targets.max(), cbar_kwargs={'label': variable_name + ' Concentration  [mmol m-2]'})
    ((variable-model) / variable * 100).plot(ax=ax[1,0], cmap=cmap, cbar_kwargs={'label': variable_name + ' Concentration  [percentage]'})

    plt.subplots_adjust(left=0.1,
        bottom=0.1, 
        right=0.95, 
        top=0.95, 
        wspace=0.35, 
        hspace=0.35)

    sa_vi.set_aspect(ax[0,0])
    sa_vi.set_aspect(ax[0,1])
    sa_vi.set_aspect(ax[1,0])


    ax[0,0].title.set_text(variable_name + ' (targets)')
    ax[0,1].title.set_text(variable_name + ' (outputs)')
    ax[1,0].title.set_text('targets - outputs')
    ax[1,1].axis('off')

    fig.suptitle(str(dates[i].date()))

    plt.show()
    

Training (Random Points)¶

In [ ]:
ds = xr.open_dataset('/data/ibougoudis/MOAD/files/integrated_model_var_old.nc')
ds2 = xr.open_dataset('/data/ibougoudis/MOAD/files/external_inputs.nc')

ds = ds.isel(time_counter = (np.arange(0, len(ds.time_counter),2)), 
    y=(np.arange(ds.y[0], ds.y[-1], 5)), 
    x=(np.arange(ds.x[0], ds.x[-1], 5)))

ds2 = ds2.isel(time_counter = (np.arange(0, len(ds2.time_counter),2)), 
    y=(np.arange(ds2.y[0], ds2.y[-1], 5)), 
    x=(np.arange(ds2.x[0], ds2.x[-1], 5)))

dates = pd.DatetimeIndex(ds['time_counter'].values)

drivers, diat, _ = datasets_preparation(ds, ds2)

regr = regressor(drivers, diat)

Other Years (Anually)¶

In [ ]:
years = range (2007,2024)

r_all = []
rms_all = []
slope_all = []

for year in tqdm(range (2007,2024)):
    
    dataset = ds.sel(time_counter=str(year))
    dataset2 = ds2.sel(time_counter=str(year))
    
    drivers, diat, _ = datasets_preparation(dataset, dataset2)

    r, rms, m = regressor2(drivers, diat, 'Diatom ' + str(year))
    
    r_all.append(r)
    rms_all.append(rms)
    slope_all.append(m)
    
plotting(np.transpose(r_all), 'Correlation Coefficient')
plotting(np.transpose(rms_all), 'Root Mean Square Error')
plotting (np.transpose(slope_all), 'Slope of the best fitting line')
  0%|          | 0/17 [00:00<?, ?it/s]
The amount of data points is 70794
The slope of the best fitting line is  0.738
The correlation coefficient is: 0.918
 The mean square error is: 0.06581317458970341
No description has been provided for this image
  6%|▌         | 1/17 [01:16<20:22, 76.39s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.729
The correlation coefficient is: 0.883
 The mean square error is: 0.06867923153926321
No description has been provided for this image
 12%|█▏        | 2/17 [02:27<18:16, 73.09s/it]
The amount of data points is 68931
The slope of the best fitting line is  0.756
The correlation coefficient is: 0.927
 The mean square error is: 0.07751808111193076
No description has been provided for this image
 18%|█▊        | 3/17 [03:36<16:38, 71.35s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.726
The correlation coefficient is: 0.915
 The mean square error is: 0.0619161166591015
No description has been provided for this image
 24%|██▎       | 4/17 [04:43<15:04, 69.54s/it]
The amount of data points is 68931
The slope of the best fitting line is  0.806
The correlation coefficient is: 0.932
 The mean square error is: 0.05813325524217016
No description has been provided for this image
 29%|██▉       | 5/17 [05:50<13:42, 68.56s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.777
The correlation coefficient is: 0.931
 The mean square error is: 0.060037348624317505
No description has been provided for this image
 35%|███▌      | 6/17 [06:55<12:24, 67.65s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.718
The correlation coefficient is: 0.915
 The mean square error is: 0.07707629395431721
No description has been provided for this image
 41%|████      | 7/17 [08:02<11:12, 67.29s/it]
The amount of data points is 68931
The slope of the best fitting line is  0.739
The correlation coefficient is: 0.895
 The mean square error is: 0.06589118551422224
No description has been provided for this image
 47%|████▋     | 8/17 [09:08<10:01, 66.82s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.741
The correlation coefficient is: 0.926
 The mean square error is: 0.06057874111103106
No description has been provided for this image
 53%|█████▎    | 9/17 [10:15<08:56, 67.08s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.795
The correlation coefficient is: 0.943
 The mean square error is: 0.05847706866111131
No description has been provided for this image
 59%|█████▉    | 10/17 [11:25<07:55, 67.98s/it]
The amount of data points is 68931
The slope of the best fitting line is  0.768
The correlation coefficient is: 0.914
 The mean square error is: 0.05530045261979468
No description has been provided for this image
 65%|██████▍   | 11/17 [12:35<06:51, 68.56s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.688
The correlation coefficient is: 0.887
 The mean square error is: 0.07499811102953335
No description has been provided for this image
 71%|███████   | 12/17 [13:46<05:45, 69.07s/it]
The amount of data points is 68931
The slope of the best fitting line is  0.698
The correlation coefficient is: 0.886
 The mean square error is: 0.08225232919417001
No description has been provided for this image
 76%|███████▋  | 13/17 [14:56<04:37, 69.38s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.697
The correlation coefficient is: 0.908
 The mean square error is: 0.09058791296148527
No description has been provided for this image
 82%|████████▏ | 14/17 [16:05<03:28, 69.42s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.801
The correlation coefficient is: 0.935
 The mean square error is: 0.0629141513700581
No description has been provided for this image
 88%|████████▊ | 15/17 [17:14<02:18, 69.29s/it]
The amount of data points is 68931
The slope of the best fitting line is  0.749
The correlation coefficient is: 0.908
 The mean square error is: 0.06267898382437069
No description has been provided for this image
 94%|█████████▍| 16/17 [18:24<01:09, 69.34s/it]
The amount of data points is 70794
The slope of the best fitting line is  0.698
The correlation coefficient is: 0.897
 The mean square error is: 0.0762321562720096
No description has been provided for this image
100%|██████████| 17/17 [19:33<00:00, 69.00s/it]
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image

Other Years (Daily)¶

In [ ]:
r_all2 = np.array([])
rms_all2 = np.array([])
slope_all2 = np.array([])

for i in tqdm(range (0, len(ds.time_counter))):
    
    dataset = ds.isel(time_counter=i)
    dataset2 = ds2.isel(time_counter=i)

    drivers, diat, _ = datasets_preparation(dataset, dataset2)

    r, rms, m = regressor3(drivers, diat)

    r_all2 = np.append(r_all2,r)
    rms_all2 = np.append(rms_all2,rms)
    slope_all2 = np.append(slope_all2,m)

plotting2(r_all2, 'Correlation Coefficients')
plotting2(rms_all2, 'Root Mean Square Errors')
plotting2(slope_all2, 'Slope of the best fitting line')
100%|██████████| 640/640 [11:11:27<00:00, 62.95s/it]  
No description has been provided for this image
No description has been provided for this image
No description has been provided for this image

Daily Maps¶

In [ ]:
maps = random.sample(range(0,len(ds.time_counter)),10)

for i in tqdm(maps):

    dataset = ds.isel(time_counter=i)
    dataset2 = ds2.isel(time_counter=i)
    drivers, diat, indx = datasets_preparation(dataset, dataset2)

    diat_i = dataset['Diatom']

    regressor4(drivers, diat, 'Diatom ')
  0%|          | 0/10 [00:00<?, ?it/s]
The amount of data points is 1863
The slope of the best fitting line is  0.647
The correlation coefficient is: 0.897
 The mean square error is: 0.10480773003649026
No description has been provided for this image
No description has been provided for this image
 10%|█         | 1/10 [01:05<09:52, 65.80s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.606
The correlation coefficient is: 0.773
 The mean square error is: 0.06227340657842682
No description has been provided for this image
No description has been provided for this image
 20%|██        | 2/10 [02:11<08:45, 65.70s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.963
The correlation coefficient is: 0.853
 The mean square error is: 0.03875772320600051
No description has been provided for this image
No description has been provided for this image
 30%|███       | 3/10 [03:18<07:43, 66.24s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.451
The correlation coefficient is: 0.548
 The mean square error is: 0.10725482913764549
No description has been provided for this image
No description has been provided for this image
 40%|████      | 4/10 [04:27<06:43, 67.32s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.595
The correlation coefficient is: 0.854
 The mean square error is: 0.07082650334617817
No description has been provided for this image
No description has been provided for this image
 50%|█████     | 5/10 [05:35<05:37, 67.51s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.67
The correlation coefficient is: 0.796
 The mean square error is: 0.06617496392247635
No description has been provided for this image
No description has been provided for this image
 60%|██████    | 6/10 [06:44<04:31, 67.97s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.905
The correlation coefficient is: 0.921
 The mean square error is: 0.02139780378862054
No description has been provided for this image
No description has been provided for this image
 70%|███████   | 7/10 [07:55<03:26, 68.98s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.661
The correlation coefficient is: 0.865
 The mean square error is: 0.06785533182710528
No description has been provided for this image
No description has been provided for this image
 80%|████████  | 8/10 [09:02<02:16, 68.41s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.719
The correlation coefficient is: 0.856
 The mean square error is: 0.06970611692619857
No description has been provided for this image
No description has been provided for this image
 90%|█████████ | 9/10 [10:09<01:08, 68.08s/it]
The amount of data points is 1863
The slope of the best fitting line is  0.683
The correlation coefficient is: 0.902
 The mean square error is: 0.07775188441965246
No description has been provided for this image
No description has been provided for this image
100%|██████████| 10/10 [11:17<00:00, 67.73s/it]
100%|██████████| 10/10 [11:17<00:00, 67.73s/it]
In [ ]: